Number 2 Volume 13 June 2007 Journal of Engineering

NATURAL CONVECTION HEAT TRANSFER IN A VERTICAL

CONCENTRIC ANNULUS

Dr. Akeel A. Mohammed

Mech. Engr. Dept.
University of Technology
Baghdad-Iraq

ABSRACT
Experiments were carried out to study the local and average heat transfer by natural convection in
a vertical concentric cylindrical annulus. The experimental setup consists of an annulus has a radius ratio
of 0.555 and inner cylinder with a heated length 1.2m subjected to the constant heat flux while the outer
cylinder is subjected to the ambient temperature. The investigation covers heat flux range from 58.2
W/m2 to 274.31 W/m2. Results show an increase in the natural convection as heat flux increases leads to
an improve in the heat transfer process. An empirical equation of average Nusselt number as a function of
Raylieh number was deduced .

‫ا‬
  ‫ا" !  ا
آ‬#‫  " أ‬$ %  ‫أ  رب 
 را ا ل ا ارة ا
)' و ا
'ل  
& ا‬
ّ
< ، 7 ‫اري‬$ 89    :‫ا اا‬#;‫ =< ا‬. ! 1.2 ‫ل‬# ‫ و‬0.555 ‫
' دل‬, #‫ ا‬%- ./ ،‫ا'
دي‬
<ّ . 274.31 W/m2 ?‫ إ‬58.2 W/m2 "! ‫ ا اري اوح‬8B  ‫ ارا !ى‬#D .‫ارة ا‬$ ‫ا ا= ر إ? در‬#;‫ا‬
.‫" !' !& ا ل ا ارة‬/  ?‫دي إ‬J
! ‫ ا اري‬8B‫ت ا
& ا   دة ا‬H'! ‫ ز دة‬EF <‫ا‬

KEY WORDS
Heat Transfer, Natural Convection, Concentric Annulus

1417
 A.l A. Mohammed Natural Convection Heat Transfer in A vertical
Concentric Annulus

INTRODUCTION
The problem of natural convection heat transfer across a horizontal and vertical cylindrical
annulus has received considerable attention in view of its fundamental importance germane to numerous
engineering applications. As a result, extensive experimental and theoretical works dealing with the flow
and associated heat transfer characteristics of natural convection in such configuration have been reported
in the literature. Comprehensive reviews on natural convection in concentric and eccentric annuli are
available (Kueehn 1976 &1978, Van de Sande 1979, and Yao 1980) and there is no need to repeat them.
However, all of the previous studies are concerned with the horizontal annulus; little attention has been
paid to annuli with a vertical position. (Van de Sande and Hamer 1979) have obtained empirical
correlations for natural convection heat transfer in concentric and eccentric annuli of constant heat flux.
(Glakape et al. 1986) presented a numerical solution for air in concentric and eccentric configurations
with specified constant heat flux at the boundaries. (Akeel 2005) has presented an experimental and
theoretical study for mixed convection heat transfer through concentric annuli. The lack of experimental
data concerning the problem of natural convection in the vertical concentric annulus, and the practical
importance of this problem in the industry applications, motivated the present work.

EXPERIMENTAL APPARATUS
The test section is shown diagrammatically in Fig.1 and consists of 4 mm wall thickness, 50 mm
outside diameter and 1.2 m long aluminum cylinder (K) located centrally in 5 mm thickness , 90 mm
inside diameter and 1.2 m long aluminum cylinder (I), by fitting it at the test section inlet with the 20 mm
inside diameter , 50 mm outside diameter and 15 mm long Teflon tube (N) and at the test section exit
with the teflon piece (M). A ring (P) is used to hold and support the aluminum cylinder (K) with the
teflon piece (N) centrally inside the settling chamber by adjustable screws (Q). The teflon was chosen
because of its low thermal conductivity in order to reduce the heat loss from the aluminum cylinder ends.
A well design Teflon bell mouth (H) is fitted at the annulus outer aluminum cylinder (I) and bolted inside
the settling chamber (D). The inlet air temperature was measured by one thermocouple (J) located in the
settling chamber (D) while the outlet bulk air temperature was measured by three thermocouples (Z)
located in the test section exit . The local bulk air temperature was calculated by using a straight line
interpolation between the measured inlet and outlet bulk air temperature.

The inner cylinder was heated electrically using an electrical heater which consists of a nickel-
chrome wire , wound as a coil spirals around solid teflon tube and is covered by a 2 mm thickness

1418
 Number 2 Volume 13 June 2007 Journal of Engineering

asbestos layer , and the space between the asbestos and the inner cylinder wall is fitted with a fine grade
sand to avoid heat convection in it and to smooth out any irregularities in the heat flux. The hole
apparatus is designed with a view to obtain a good concentricity of the core cylinder and the containing
cylinder. The temperature of the outside surface of the inner cylinder was measured by seventeen asbestos
sheath alumel-chromel (type K) thermocouples , arranged along the cylinder , the measuring heads of the
thermocouples were made by fusing together the ends of two wires.

The thermocouples were fixed by drilling holes of 1.5 mm diameter in the cylinder wall and the
ends of the holes chamfered by a 3 mm slug to locate the measuring junctions which were then fixed by a
high temperature application Defcon adhesive . The excess adhesive was removed and the cylinder outer
surface was cleaned carefully by fine grinding paper. All the thermocouples wires and heater terminals
were taken out the test section through both teflon pieces (N,M) .
On the other hand , ten thermocouples (type K) were used to measure the inner surface
temperature of the annulus outer cylinder (I). Thermocouples positions at the outer surface were located
and then a 2 mm deep pits were drilled in which the thermocouples were fixed by Defcon adhesive. All
thermocouples were used with leads , the thermocouple with lead and without lead were calibrated
against the melting point of ice made from distilled water and the boiling points of several pure chemical
substances. To determine the heat loss from the test section ends, two thermocouples were fixed in each
teflon piece. The distance between these thermocouple was 12 mm. Knowing the thermal conductivity of
the teflon , the ends condition could thus be calculated.

EXPERIMENTAL PROCEDURE
To carry out an experiment the following procedure was followed:

1- The electrical heater was switched on and the heater input power then adjusted to give the required
heat flux.
2- The apparatus was left at least three hours to establish steady state condition. The thermocouples
readings were measured every half an hour by means of the digital electronic multimeter until the reading
became constant , a final reading was recorded. The input power to the heater could be increased to cover
another run in a shorter period of time and to obtain steady state conditions for next heat flux.
3- During each test run , the following readings were recorded:
a- The readings of the thermocouples in ˚C.
b- The heater current in amperes.
c- The heater voltage in volts.

1419
 A.l A. Mohammed Natural Convection Heat Transfer in A vertical
Concentric Annulus

DATA ANALYSIS
Simplified steps were used to analyze the heat transfer process by natural convection from the inner
cylinder which was subjected to a uniform heat flux while the outer cylinder was subjected to the ambient
temperature. The total input power supplied to the inner cylinder can be calculated:

Qt=V” ×I (1)

The convection and radiation heat transferred from the inner cylinder is :

Qcr=Qt- Qcond (2)

where Qcond is the conduction heat loss which was found experimentally equal to 5 % of the input power.
The convection and radiation heat flux can be represented by:

qcr= Qcr/A (3)

where:
A = 2π r1 L
The convection heat flux , which is used to calculate the local heat transfer coefficient is obtained
after deduce the radiation heat flux from qcr value. The local radiation heat flux can be calculated as
follows:

q
r
=F
1 − 2
[(
ε σ (t ) + 273
s z
) (
4
− (t )
s2 z
+ 273 )]
4
(4)

where:
F1-2= view factor between inner and outer cylinder ≈1

(t )z = local temperature of inner cylinder.

s

1420
 Number 2 Volume 13 June 2007 Journal of Engineering

(t ) = average temperature of outer cylinder.

s2 z

ε = emissivity of the polished aluminum surface=0.09.

Hence the convection heat flux at any position is:

q = qcr – qr (5)

The local heat transfer coefficient can be obtained as:

q
hz = (6)
(ts )z − (t b )z

(t b )z = Local bulk air temperature.

All the air properties were evaluated at the mean film air temperature (Keys 1966):

(t )z + ( t )
(t ) = s b z (7)
f z 2

t f = Local mean film air temperature.

The local Nusselt number (Nuz) then can be determine as:

h Dh
Nuz = z (8)
κ

The average values of Nusselt number Num can be calculated based on calculation of average inner
surface temperature and average bulk air temperature as follows:

1 z =L
(9)
s L z =∫ 0
t = (t )z dz
s

1421
 A.l A. Mohammed Natural Convection Heat Transfer in A vertical
Concentric Annulus

1 z =L
b L z =∫ 0 b z
t = (t ) dz (10)

tz + tb
t = (11)
f 2

q D
Num = h (12)
k(t − t b )
z

The average values of the other parameters can be calculated as:

Grm =
g β Dh 3 (t z − t b ) (13)
ν2

µ Cp
Prm = (14)
k

Ram=Grm . Prm (15)

where:

(
β = 1 273 + t
f
)
All the air physical properties ρ, µ, ν, and k were evaluated at the average mean film temperature

( t ).
f

EXPERIMENTAL RESULTS
The variation of the inner cylinder surface temperature for different heat flux is shown in Fig.2 . It
is obvious that the surface temperature increases at the stage of entrance and attains a maximum point
after which the surface temperature begins to decrease at high heat flux (q≥181 W/m2) and be almost
constant for small heat flux (q≤152 W/m2). The rate of surface temperature rises at early stage is directly

1422
 Number 2 Volume 13 June 2007 Journal of Engineering

proportional to the wall heat flux because of the faster increasing of the thermal boundary layer as heat
flux increase (i.e., increasing of buoyancy effect).
Fig.3 & 4 show the effect of heat flux on the local and average Nusselt number along the inner
cylinder. It is clear that the results of higher heat flux for local and average Nusselt number are higher
than that of lower heat flux. Fig.3 shows also sharp decrease for the local Nusselt number values at the
entrance of the annulus because the boundary layer thickness is zero and the natural convection is poor in
this region, then increase downstream because of increasing of natural convection. The values of the
mean Nusselt number are plotted in Fig.5 in the form of log(Num) against log(Ra) for the range of Ra
from 0.68611×105 to 1.728559×105. All the points as can be seen are represented by linearization of the
following equation.

Num=2.31812 Ra0.083188 (16)

CONCLUSIONS
1. The extent of the local mixing increases as the heat flux increases.
2. The heat transfer process improves as heat flux increase.
3. The effect of buoyancy is small at the annulus entrance and increase down stream.

REFERNCES
Akeel Abdullah Mohammed An Investigation into laminar combined convection heat transfer through
concentric annuli. Ph. D. Thesis, University of Technology, Mech. Eng. Dept., 2005.
Glakape, E. K., Watkins C. B. , Jr., and Cannon, J. N. Constant heat flux solutions for natural convection
between concentric and eccentric horizontal cylinders. Numerical Heat Transfer, 1986, 10, 279-295
Kueehn, T. H. and Goldstein, R. J. An experimental and theoretical study of natural convection in the
annulus between horizontal concentric clinders. J. Fluid Mech., 1976, 74, 695-719
Kueehn, T. H. and Goldstein, R. J. An experimental study of natural convection heat transfer in
concentric and eccentric horizontal cylindrical annuli, ASME J. Heat Transfer, 1978, 100, 635-640
Rao, Y. F., Miki,K. , Fukuda, K. , and Katata, Y. Flow pattern of natural convection in horizontal
cylindrical annuli. Int. J. Heat Mass Transfer, 1985, 28, 705-714
Van de Sande, E. and Hamer, B. J. G. Steady and transient natural convection in enclosures between
horizontal circular cylinders (constant heat flux). Int. J. Heat Mass Transfer, 1979, 22, 361-370

1423
 A.l A. Mohammed Natural Convection Heat Transfer in A vertical
Concentric Annulus

Yao, L. S. Analysis of heat transfer in slightly eccentric annuli. ASME J. Heat Transfer, 1980, 102, 270-
284
NOMENCLUTURE
A: inner cylinder surface area; m2
Dh: hydraulic diameter=2(r2-r1): m
I: current; Amp
κ : thermal conductivity; W/m2.°C
L: annulus length; m
(Num): mean Nusselt number
Q: convection heat loss; W
Qt: total heat given; W
Qcr: convection- radiation heat loss; W
qr: radiation heat flux; W/m2.°C
q: convection heat flux; W/m2.°C
r1: outer radius of inner cylinder; m
r2: inner radius of outer cylinder; m
V”: voltage; volt

1424
Number 2 Volume 13 June 2007 Journal of Engineering

A
Q D Q
P N M Z
H
I
J K M

A WOODEN BOARD

D SETTLING CHAMPER

H BELL MOUTH

I OUTER CYLINDER

K INNER CYLINDER

J THERMOCOUPLE

M,N TEFLON

P RING

Z THERMOCOUPLE

Q SCREW

Fig. 1: Diagram of Experimental Apparatus.

1425
A.l A. Mohammed Natural Convection Heat Transfer in A vertical
Concentric Annulus

160
2
q=58.20 W/m q=181 W/m2
150 2
=81.73 W/m2 =229 W/m
140 =120.33 W/m2 =274.31 W/m2
130 =152.84 W/m2

120

110
t( C)

o 100

90

80

70

60

50

40
0 0.2 0.4 0.6 0.8 1 1.2
z(m)

Fig.2: Variation of the Surface Temperature with the Axial Distance

10
2
q=58.20 W/m q=181 W/m2
2
=81.73 W/m2 =229 W/m
9
=120.33 W/m2 =274.31 W/m2
=152.84 W/m2

7
Nuz

4
0.1 1
z(m)

Fig.3: Local Nusselt Number Versus Axial Distance.

1426
Number 2 Volume 13 June 2007 Journal of Engineering

10
2
q=58.20 W/m
9
=81.73 W/m2
=181 W/m2
=229 W/m2
8 =274.31 W/m2
Num

4
0.1 1
z(m)

Fig.4: Average Nusselt Number Versus Axial Distance.

0.083188
Nu m =2.31812 Ra
Log(Nu m )

1E+5
Log(Ra)

Fig.5: Logarithm Average Nusselt Number Versus log(Ra).

1427